Maintaining a large network of server and client machines is a daunting task, For example, consider a company having 500 Linux based bar-code scanning devices (client machines) used at various locations around the world. A technician is responsible for monitoring the output of the client machines. The technician sees that a specific client machine is causing errors in the system and would like to connect to that client machine. However, the technician cannot initiate a connection to the client machine because it is within a network with fire-walls preventing the connection. Using conventional systems and methods, the technician could maintain an open connection between all the client machines (500 or more) and the server but this would take bandwidth and system resources. Alternatively, a conventional ICMP request could be sent but that would either cause each client machine to constantly ping the server machine or there would be a delay in the connection. Still alternatively, an http request could be used to poll the server machine but that would also increase the bandwidth use and resources.